Issue #33: My Generation

🎶 I’m not trying to cause a big sensation, I’m just talkin’ ‘bout my generation… 🎶

The last fortnight brought a lot of mind-boggling updates of all sorts. Starting with framework-level philosophical doctrines and up to Oscar-worthy performances from JS bundler developers, with some habitual controversies, and many highly-anticipated releases. Metaframework developers give abundant food for thought almost daily, so let’s enjoy the spicy snack. Bon appétit!

The Good

October is a conference season, especially in the JS world, and in-person communication brings a lot of inspiration to event participants, along with the true sense of community. This is the time when initiatives like Svelte tenets surface and find genuine support, especially against the background of the recent dubious events.

As a community our default position is one of optimism about technology — the platform is getting better, our tools are getting better, our devices are getting better, and if we embrace that fact we can make better stuff.

And one of the main heroes of ecosystem enbetterment is Vite. The recent (first in-person!) ViteConf brought up a whole bunch of goodies from the team of maintainers and the community, with the most heartwarming one being the new Vite documentary.

If you’re using a JavaScript framework, you’re probably using Vite.

Evan You announced early access for Vite+, the unified toolchain for web developers, explaining how the VoidZero team is going to monetize all the titanic and Rust-heavy surrounding efforts. It would be fair to assume the team was inspired by the UmiJS ecosystem.

Another important announcement was JavaScript plugins support in Oxlint, the blazingly-fast ESLint killer. As the new plugins API is compatible with ESLint’s plugins, migrating to Oxlint may become a no-brainer for people who don’t like linter-run-waiting-driven coffee breaks.

Vitest also got its own star moment with the version 4 announcement, bringing improvements that may make you think about getting rid of most of your other testing-related dependencies (with browser mode, visual regression testing, and Playwright traces for better test debugging).

But not only VoidZero had something to brag about. Vercel announced beta for Next 16 (I have just started to get used to the number 15!), and published the main release almost right after that. The team has a good answer to Vite — Turbopack, which is now declared to be stable, alongside support for the latest and greatest from the React world (including compiler) and improvements to partial prerendering. Also, the new proxy file replaces a middleware file, probably hoping adversaries won’t spot middleware vulnerabilities anymore, heh.

The last of the comparatively huge controversial news from the latest weeks was the introduction of Remix 3, or rather some new details on what it’s going to be, which drew a lot of attention from the metaframeworks-savvy community and React fans (and haters alike) — though as it happens, there’s neither React nor even Preact in this whole story. The whole idea would be much more obscure if not for the cool guy Mark Dalgleish and his Remix 3 Resources repo, where you can find both theoretical details and practical examples of Remix 3 usage. You can even use it with Storybook already, which is quite exciting for those who like a good story.

The Bad

And while new metaframeworks come, some old friends do go away. One of the pioneers of the metaframework+ (a metaframework with blackjack and hookers more bells and whistles) movement, RedwoodJS, has got its last update and was announced to no longer be maintained, in favor of RedwoodSDK. You can also use the less-known fork of it — CedarJS (or use other full-blown alternatives like Wasp or Blitz.js), but this nostalgic feeling of a Thomas Preston-Werner epoch fading away is inevitable.

Among other bad news of the latest weeks was the security breach found in Better-auth — one of the most popular framework-agnostic authentication tools used in the JS ecosystem. With authentication being the least acceptable part of your app to give up to adversaries, this account-takeover breach brings doubts in the ubiquitous mantra of “always use existing solutions for everything”.

But maybe it’s just a matter of good security solutions we developers should be aware of. And while I’m not a big fan of SDK-driven approaches in this regard, tools like Arcjet, which provides support for Nuxt and React Router now, might be a minimal step toward avoiding the threats we all have already stopped ignoring, especially lately.

The Noteworthy

While Vite is taking the industry by storm on more and more levels, starting historically from the top (front end), Bun goes up and up from their low-level runtime ground, widening the range of tools they provide. Version 1.3 brings a lot of decent updates, including full-stack dev server, more support for SQL DBs, integrated Redis client, and traditional (and important) Node compatibility improvements. And even though some people doubt the project is ready yet, who knows, maybe someday we won’t need [meta]frameworks at all, just open Bun and start the full-stack webapp ride…

But until we’re there, I’m still excited about (even the early) releases of tools like Nitro, whose version 3 is coming closer with lots of long-awaited updates that more and more metaframeworks share these days.

Another important pre-release of recent times in the JS ecosystem was the beta of Preact 11, with which the team is going to refresh the framework while avoiding drastic breaking changes as much as possible at the same time.

Something that brings some hope for developing original innovations in the industry is the return of Aral Roca with version 0.2.15 of his Bun-bound Brisa metaframework.

And last but not least — important middleware improvements in version 4.10.0 of Hono.

A lot of stuff changes constantly in the metaframeworks world, but one thing remains — the permanent striving for innovation in a diverse way, which is the most important thing, as this is the only path to this technology becoming a boring technology eventually — in a good meaning of this term. Meanwhile, I’m happy my generation of metaframeworks developers don’t sit idle but search for new awesome ways to create full-stack web applications faster and better. Here’s to more records on this journey!

🥂